PINNING DOWN POTENTIAL 21In other words, humans share the vast majority of their genes and their
nucleotide sequences: any two humans taken at random from anywhere
in the world will be more than 99 percent identical. On one hand, ten mil-
lion SNPs still represent lots of variation. On the other hand, we also
know that nearly all these variations are functionally neutral: it doesn’t
matter which version you have, they work equally well (as I explain in
chapter 4). Trying to segregate those relatively few SNPs that supposedly
make a diff erence from those that do not is diffi cult enough for a well-
defi ned medical condition. But for traits as poorly defi ned as intelligence
and using only statistical correlations as evidence, making fi rm conclu-
sions already seems highly naïve.
In fact, the method of associating such tiny pieces of DNA with crude
and imprecise mental scores— oft en obtained from shortened tests admin-
istered over the telephone, online, or by parents to their children, and
usually statistically compressed (see chapter 2)—is almost designed for
churning out reams of spurious correlations. And that is what is happen-
ing. Th e statistical associations seem to be scintillations in a fog.
An impor tant part of the fog is that complex physiological and mental
functions do not vary like eye or hair color, through a single gene or very
few genes. Educational and cognitive “phenotypes” are not simple physi-
cal traits, like height and weight, milk yield in cows, or egg laying in hens.
Th ey involve thousands or even tens of thousands of genes. And they are
not utilized like electrical charges being added together. Instead, individ-
ual genes are taken up as resources into complex biochemical networks
that create variation out of intense interactions.
Most of the fog, however, is around what is said to be mea sured.
Most reported correlations are with some sort of IQ scores. But there
is little agreement about what they really mea sure except in equally ob-
scure terms. Th e Imperial College group suggests that they have turned
cutting- edge gene- identifi cation technology to fi nding genes related
to memory, attention, pro cessing speed, reasoning, and executive func-
tion. Th e claim is made with great confi dence, as if those were clearly
defi ned and characterized functions, like (some) medical conditions.
Th ey are emphatically not, as I explain in chapters 3 and 7. Th e best anal-
ogy I can think of is using super high- tech surveillance equipment to
capture a creature in the woods in the dark when we aren’t sure what it
This content downloaded from 139.184.14.159 on Tue, 17 Oct 2017 13:51:16 UTC